Video Abstract

Video Abstract

BACKGROUND:

Available evidence supports ovary-sparing surgery for benign ovarian neoplasms; however, preoperative risk stratification of pediatric ovarian masses can be difficult. Our objective of this study was to characterize the surgical management of pediatric ovarian neoplasms across 10 children’s hospitals and to identify factors that could potentially aid in the preoperative risk stratification of these lesions.

METHODS:

A retrospective review of girls and women aged 2 to 21 years who underwent surgery for an ovarian neoplasm between 2010 and 2016 at 10 children’s hospitals was performed. Multivariable logistic regression was used to examine the relationships between the preoperative cohort characteristics, procedure performed, and risk of malignancy.

RESULTS:

Among 819 girls and women undergoing surgery for an ovarian neoplasm, malignant lesions were identified in 11%. The overall oophorectomy rate for benign disease was 33% (range: 15%–49%) across institutions. Oophorectomy for benign lesions was independently associated with provider specialty (P = .002: adult gynecologist, 45%; pediatric surgeon, 32%; pediatric gynecologist, 18%), premenarchal status (P = .02), preoperative suspicion for malignancy (P < .0001), larger lesion size (P < .0001), and presence of solid components (P < .0001). Preoperative findings independently associated with malignancy included increasing size (P < .0001), solid components (P = .003), and age (P < .0001).

CONCLUSIONS:

The rate of oophorectomy for benign ovarian disease remains high within the pediatric population. Identification of factors associated with the choice of procedure and the risk of malignancy may allow for improved preoperative risk stratification and fewer unnecessary oophorectomies. These results have been used to develop and validate a multidisciplinary preoperative risk stratification algorithm that is currently being studied prospectively across 10 institutions.

What’s Known on This Subject:

Ovarian neoplasms in the pediatric population have a low rate of malignancy. There is still substantial practice variation in the management of ovarian neoplasms, with patients potentially undergoing unnecessary oophorectomies for benign disease.

What This Study Adds:

Oophorectomy for benign disease is performed frequently with variation by site and specialty. This study identified factors associated with the choice of procedure and the risk of malignancy that may allow for improved preoperative risk stratification and fewer unnecessary oophorectomies.

Ovarian masses in children are a relatively rare condition, with an incidence of 2.2 to 2.6 cases per 100 000 female patients each year.1,2 Approximately 75% to 90% of these neoplasms are benign, with the remaining 10% to 25% being either malignant tumors or tumors with malignant potential.2,5 Depending on their age and access to specialists, patients with an ovarian mass may be cared for by pediatric surgeons, pediatric and adolescent gynecologists (PAGs), or adult gynecologists. Previous studies have revealed that management of ovarian masses varies by both institution and surgeon specialty.6,8 

In the management of ovarian neoplasms, the surgeon has the option of monitoring the patient or proceeding to an operation. If the choice is made to operate, the decision must be made whether to remove just the mass (ovary-sparing surgery) or remove the mass along with the entire ovary (oophorectomy). Although the available evidence supports ovary-sparing surgery for benign ovarian neoplasms, recent studies have revealed that pediatric patients with a benign mass undergo an oophorectomy in up to 50% of cases.6,9,10 

One of the major challenges in approaching the pediatric patient with an ovarian mass is the preoperative stratification of benign versus malignant disease, a distinction that plays a major role in operative decision-making.11 Our objective of this study was to characterize the surgical management of pediatric ovarian neoplasms across 10 children’s hospitals and identify factors that could potentially aid in the preoperative risk stratification of these lesions.

We conducted a multi-institutional retrospective cohort study of patients 2 to 21 years of age with ovarian neoplasms who underwent surgical management between January 1, 2010, and December 31, 2016, at 10 participating institutions of the Midwest Pediatric Surgery Consortium (www.mwpsc.org). This study was approved by the institutional review boards with a waiver of consent.

Eligible patients were identified by using the International Classification of Disease, Ninth Edition, Clinical Modification and Current Procedural Terminology procedure codes for ovarian neoplasms (see Supplemental Tables 5 and 6). Female patients aged 2 to 21 years with an ovarian neoplasm undergoing surgical management were included. Patients with simple cysts, congenital ovarian abnormalities, torsion without a neoplasm, or with unavailable pathology results were excluded. The charts were reviewed for patient characteristics, admission characteristics, laboratory and imaging results, operative findings, pathology reports, and postoperative complications. The data were recorded and managed by using the Research Electronic Data Capture tool.

The 2 primary objectives for this study were (1) to identify risk factors for malignant ovarian neoplasms in this population (on the basis of postoperative pathology reports) and (2) to identify characteristics associated with oophorectomies for benign neoplasms.

Patient demographic characteristics, whether the case was elective, specialty of the operating physician (pediatric surgeon, PAG, adult gynecologist, or other specialty), year of treatment, preoperative suspicion of malignancy, and laterality were evaluated as risk factors for the outcomes. Also evaluated were imaging characteristics, such as lesion size, presence of septations, free fluid, solid components, torsion, and papillary projections, and recorded patient symptoms from the medical record. Data on preoperative tumor markers (β–human chorionic gonadotropin, lactate dehydrogenase, α-fetoprotein (AFP), cancer antigen 125, cancer antigen 19-9, carcinoembryonic antigen, progesterone, inhibin A, or inhibin B) were collected but were not included in the analysis because testing was highly variable and inconsistent (range: 52% had AFP collected and 1% had progesterone levels).

Frequencies and percentages as well as medians and interquartile ranges (IQRs) were used to describe categorical and continuous variables, respectively. Bivariate relationships were assessed by using χ2, Fisher’s exact, and Wilcoxon–Mann–Whitney U tests, where appropriate. The 2 dependent variables, malignancy and oophorectomy for benign neoplasms, were analyzed by using logistic mixed-effects models with hospital-level random intercepts to account for clustering by hospital. Missing data were treated with dummy variables to avoid the exclusion of additional participants. Patients treated by an “other provider” (n = 4) and those missing data on lesion size (n = 115) were excluded from multivariable analyses. Differences in case mix by provider specialty were assessed. Interactions between provider specialty and clinical characteristics that were significantly different across provider specialty were included in our multivariable model of oophorectomy for benign disease. Multivariable models included all variables with a P < .05 in bivariate analyses and all interactions with P < .1. All statistical analyses were performed by using SAS Enterprise Guide version 7.1 (SAS Institute, Inc, Cary, NC). Statistical significance was defined as P < .05.

Chart review was performed for 819 patients who underwent surgical management for an ovarian neoplasm (Table 1). The median age was 14.6 years (IQR: 12.1, 16.7), and 63% of patients were non-Hispanic white individuals. The most common surgical procedures performed were ovarian cystectomy and salpingo-oophorectomy. Of the 725 benign lesions, the most common pathologic diagnoses were mature cystic teratoma (46%) and serous cystadenoma (15%). Of the 94 malignant lesions, the most common pathologic diagnoses were sex cord–stromal tumor (31%), immature teratoma (23%), and dysgerminoma (10%). Overall, 20% of patients presented with torsion, with no difference in size of the mass in patients with torsion and those without torsion (9.95 vs 10.58 cm). Tumor spillage occurred in 94 cases, occurring in 17% of malignant neoplasms and 11% of benign neoplasms; it happened during 14% of ovary-sparing surgeries and 7% of oophorectomies. Tumor spillage did not differ significantly by specialty (P = .14), with PAGs performing 36% of the cases with tumor spillage (n = 34), pediatric surgeons 40% (n = 38), and adult gynecologists 22% (n = 21).

TABLE 1

Patient Characteristics by Malignancy Status (N = 819)

CharacteristicMalignant (N = 94)Benign (N = 725)P
Age, median (IQR), y 13.8 (10.5–15.8) 14.6 (12.4–16.9) .003 
Race and/or ethnicity, n (%)   .17 
 Non-Hispanic white 65 (69.1) 449 (61.9) — 
 Non-Hispanic African American 6 (6.4) 113 (15.6) — 
 Hispanic 9 (9.6) 78 (10.8) — 
 Other and multiracial 7 (7.4) 44 (6.1) — 
 Unknown 7 (7.4) 41 (5.7) — 
Menarchal status, n (%)   .02 
 Premenarche 30 (31.9) 148 (20.4) — 
 Postmenarche 50 (53.2) 488 (67.3) — 
 Not reported 14 (14.9) 89 (12.3) — 
Patient history of previous lesion, n (%) 6 (6.4) 38 (5.2) .84 
Symptoms, n (%)    
 Abdominal pain 57 (60.6) 544 (75.0) .003 
 Nausea 20 (21.3) 145 (20.0) .77 
 Vomiting 15 (16.0) 134 (18.5) 0.55 
 Fever 5 (5.3) 28 (3.9) .50 
 Abdominal bloating 21 (23.4) 27 (3.7) <.001 
 Constipation 9 (9.6) 105 (14.5) .20 
 Palpable mass 2 (2.1) 14 (1.9) .90 
Imaging, n (%)    
 Septations 27 (28.7) 163 (22.5) .05 
 Free fluid 47 (50.0) 228 (31.4) .001 
 Solid components 69 (73.4) 324 (44.7) <.001 
 Papillary projections 2 (2.1) 1 (0.1) .002 
 Ill-defined borders 7 (7.4) 11 (1.5) <.001 
 Extension 7 (7.4) 10 (1.4) <.001 
 Lymphadenopathy 11 (11.7) 12 (1.7) <.001 
 Metastatic 18 (19.1) 5 (0.7) <.001 
 Complex 50 (53.2) 211 (29.1) <.001 
Size of mass, median (IQR), cm 14.4 (10.6–21.9) 7.6 (5.1–12.0) <.001 
Torsion noted intraoperatively, n (%) 15 (16.0) 152 (21.0) .27 
Preoperative suspicion of malignancy, n (%) 73 (77.7) 69 (9.5) <.001 
Hospital, n (%)   .003 
 1 5 (5.3) 58 (8.0) — 
 2 9 (9.6) 162 (22.3) — 
 3 14 (14.9) 48 (6.6) — 
 4 15 (16.0) 65 (9.0) — 
 5 3 (3.2) 22 (3.0) — 
 6 5 (5.3) 45 (6.2) — 
 7 5 (5.3) 88 (12.1) — 
 8 16 (17.0) 97 (13.4) — 
 9 9 (9.6) 44 (6.1) — 
 10 13 (13.8) 96 (13.2) — 
CharacteristicMalignant (N = 94)Benign (N = 725)P
Age, median (IQR), y 13.8 (10.5–15.8) 14.6 (12.4–16.9) .003 
Race and/or ethnicity, n (%)   .17 
 Non-Hispanic white 65 (69.1) 449 (61.9) — 
 Non-Hispanic African American 6 (6.4) 113 (15.6) — 
 Hispanic 9 (9.6) 78 (10.8) — 
 Other and multiracial 7 (7.4) 44 (6.1) — 
 Unknown 7 (7.4) 41 (5.7) — 
Menarchal status, n (%)   .02 
 Premenarche 30 (31.9) 148 (20.4) — 
 Postmenarche 50 (53.2) 488 (67.3) — 
 Not reported 14 (14.9) 89 (12.3) — 
Patient history of previous lesion, n (%) 6 (6.4) 38 (5.2) .84 
Symptoms, n (%)    
 Abdominal pain 57 (60.6) 544 (75.0) .003 
 Nausea 20 (21.3) 145 (20.0) .77 
 Vomiting 15 (16.0) 134 (18.5) 0.55 
 Fever 5 (5.3) 28 (3.9) .50 
 Abdominal bloating 21 (23.4) 27 (3.7) <.001 
 Constipation 9 (9.6) 105 (14.5) .20 
 Palpable mass 2 (2.1) 14 (1.9) .90 
Imaging, n (%)    
 Septations 27 (28.7) 163 (22.5) .05 
 Free fluid 47 (50.0) 228 (31.4) .001 
 Solid components 69 (73.4) 324 (44.7) <.001 
 Papillary projections 2 (2.1) 1 (0.1) .002 
 Ill-defined borders 7 (7.4) 11 (1.5) <.001 
 Extension 7 (7.4) 10 (1.4) <.001 
 Lymphadenopathy 11 (11.7) 12 (1.7) <.001 
 Metastatic 18 (19.1) 5 (0.7) <.001 
 Complex 50 (53.2) 211 (29.1) <.001 
Size of mass, median (IQR), cm 14.4 (10.6–21.9) 7.6 (5.1–12.0) <.001 
Torsion noted intraoperatively, n (%) 15 (16.0) 152 (21.0) .27 
Preoperative suspicion of malignancy, n (%) 73 (77.7) 69 (9.5) <.001 
Hospital, n (%)   .003 
 1 5 (5.3) 58 (8.0) — 
 2 9 (9.6) 162 (22.3) — 
 3 14 (14.9) 48 (6.6) — 
 4 15 (16.0) 65 (9.0) — 
 5 3 (3.2) 22 (3.0) — 
 6 5 (5.3) 45 (6.2) — 
 7 5 (5.3) 88 (12.1) — 
 8 16 (17.0) 97 (13.4) — 
 9 9 (9.6) 44 (6.1) — 
 10 13 (13.8) 96 (13.2) — 

Other race includes Asian American, American Indian or Alaska Native, Hawaian Islander or Pacific Islander, multiracial, or self-reported other. —, not applicable.

Overall, 11% of patients had malignant neoplasms, with a range of malignancy rates from 4% to 18% across the 10 hospitals. Children with malignant neoplasms were younger and more likely to be premenarche (Table 1). Imaging characteristics more commonly associated with malignancy included larger lesion size, septations, free fluid, solid components, papillary projections, ill-defined borders, extension, lymphadenopathy, and complex features (P ≤ .05 for all). The only symptom that was more common among malignant cases was abdominal bloating (P < .001).

Multivariable analyses revealed that older age was associated with a decreased odds of malignancy (P < .001) (Table 2). Similarly, abdominal bloating was associated with 2.8 times the odds of malignancy (P = .01). The presence of lymphadenopathy and solid components on imaging were associated with an increased odds of malignancy (odds ratio [OR]: 4.3, 95% confidence interval [CI]: 1.9–9.9; OR: 5.7, 95% CI: 2.2–14.4, respectively) (Table 2). Finally, each 1-cm increase in lesion size was associated with a 12% increased odds of malignancy (P < .001).

TABLE 2

Predictors of Malignant Neoplasms in a Logistic Mixed-Effects Model

PredictorsOR95% Wald Confidence LimitsP
Age, y 0.89 0.84 0.94 <.001 
Size of mass, cm 1.12 1.08 1.17 <.001 
Solid components 5.65 2.23 14.36 <.001 
Lymphadenopathy 4.33 1.90 9.87 <.001 
Abdominal bloating 2.78 1.28 6.03 .01 
PredictorsOR95% Wald Confidence LimitsP
Age, y 0.89 0.84 0.94 <.001 
Size of mass, cm 1.12 1.08 1.17 <.001 
Solid components 5.65 2.23 14.36 <.001 
Lymphadenopathy 4.33 1.90 9.87 <.001 
Abdominal bloating 2.78 1.28 6.03 .01 

Of patients with benign neoplasms, 33% underwent oophorectomy or salpingo-oophorectomy, with a range of oophorectomy rates between 15% and 49% across the 10 hospitals. Hospital volume was not related to oophorectomy rates for benign neoplasm (P = .13). The oophorectomy rate for benign neoplasms differed across provider specialty, with PAGs having the lowest rate (Table 3). Patients receiving oophorectomy were more likely to be premenarche, present with symptoms of abdominal bloating and constipation, and have preoperative suspicion of malignancy. Imaging characteristics more commonly associated with oophorectomy included larger lesion size, septations, solid components, extension, and complex features.

TABLE 3

Patient Characteristics by Procedure Among Patients With Benign Neoplasms (N = 725)

CharacteristicOvary-Sparing Surgery (N = 489)Oophorectomy (N = 236)P
Age, median (IQR), y 14.6 (12.7–16.7) 14.7 (11.4–17.0) .68 
Race and/or ethnicity, n (%)   .10 
 Non-Hispanic white 302 (61.8) 147 (62.3) — 
 Non-Hispanic African American 70 (14.3) 43 (18.2) — 
 Hispanic 61 (12.5) 17 (7.2) — 
 Other and multiracial 26 (5.3) 18 (7.6) — 
 Unknown 30 (6.1) 11 (4.7) — 
Menarchal status, n (%)   .04 
 Premenarche 90 (18.4) 58 (24.6) — 
 Postmenarche 344 (70.3) 144 (61.0) — 
 Not reported 55 (11.2) 34 (14.4) — 
Patient history of previous lesion, n (%) 29 (5.9) 9 (3.8) .38 
Symptoms, n (%)    
 Abdominal pain 390 (79.8) 154 (65.3) <.001 
 Nausea 102 (20.9) 43 (18.2) .41 
 Vomiting 89 (18.2) 45 (19.1) .78 
 Fever 15 (3.1) 13 (5.5) .11 
 Abdominal bloating 13 (2.7) 14 (5.9) .03 
 Constipation 56 (11.5) 49 (20.8) <.001 
 Palpable mass 7 (1.4) 7 (3.0) .16 
Provider specialty, n (%)   <.001 
 Pediatric adolescent gynecology 127 (26.0) 26 (11.0) — 
 Pediatric surgery 289 (59.1) 151 (64.0) — 
 Adult gynecology 71 (14.5) 57 (24.2) — 
 Other specialty 2 (0.4) 2 (0.9) — 
Elective status, n (%)   .97 
 Elective 150 (30.7) 73 (31.1) — 
 Urgent 335 (68.5) 162 (68.9) — 
 Not reported 4 (0.8) 0 (0) — 
Imaging, n (%)    
 Septations 94 (19.2) 69 (30.8) .003 
 Free fluid 149 (30.5) 79 (35.3) .34 
 Solid components 185 (37.8) 139 (62.1) <.001 
 Papillary projections 0 (0.0) 1 (0.5) .33 
 Ill-defined borders 8 (1.6) 3 (1.3) .33 
 Extension 4 (0.8) 6 (2.7) .002 
 Lymphadenopathy 8 (1.6) 4 (1.8) .28 
 Metastatic 2 (0.4) 3 (1.3) .11 
 Complex 129 (26.4) 82 (36.6) .01 
Size of mass, median (IQR), cm 6.7 (5.0–10.0) 9.7 (6.1–15.0) <.001 
Torsion noted intraoperatively, n (%) 80 (16.4) 72 (30.5) <.001 
Preoperative suspicion of malignancy, n (%) 28 (5.7) 41 (17.4) <.001 
Hospital, n (%)   <.001 
 1 43 (8.8) 15 (6.4) — 
 2 115 (23.5) 47 (19.2) — 
 3 40 (8.2) 8 (3.4) — 
 4 29 (5.9) 36 (15.3) — 
 5 15 (3.1) 7 (3.0) — 
 6 27 (5.5) 18 (7.6) — 
 7 61 (12.5) 27 (11.4) — 
 8 57 (11.7) 40 (17.0) — 
 9 30 (6.1) 14 (5.9) — 
 10 72 (14.7) 24 (10.2) — 
CharacteristicOvary-Sparing Surgery (N = 489)Oophorectomy (N = 236)P
Age, median (IQR), y 14.6 (12.7–16.7) 14.7 (11.4–17.0) .68 
Race and/or ethnicity, n (%)   .10 
 Non-Hispanic white 302 (61.8) 147 (62.3) — 
 Non-Hispanic African American 70 (14.3) 43 (18.2) — 
 Hispanic 61 (12.5) 17 (7.2) — 
 Other and multiracial 26 (5.3) 18 (7.6) — 
 Unknown 30 (6.1) 11 (4.7) — 
Menarchal status, n (%)   .04 
 Premenarche 90 (18.4) 58 (24.6) — 
 Postmenarche 344 (70.3) 144 (61.0) — 
 Not reported 55 (11.2) 34 (14.4) — 
Patient history of previous lesion, n (%) 29 (5.9) 9 (3.8) .38 
Symptoms, n (%)    
 Abdominal pain 390 (79.8) 154 (65.3) <.001 
 Nausea 102 (20.9) 43 (18.2) .41 
 Vomiting 89 (18.2) 45 (19.1) .78 
 Fever 15 (3.1) 13 (5.5) .11 
 Abdominal bloating 13 (2.7) 14 (5.9) .03 
 Constipation 56 (11.5) 49 (20.8) <.001 
 Palpable mass 7 (1.4) 7 (3.0) .16 
Provider specialty, n (%)   <.001 
 Pediatric adolescent gynecology 127 (26.0) 26 (11.0) — 
 Pediatric surgery 289 (59.1) 151 (64.0) — 
 Adult gynecology 71 (14.5) 57 (24.2) — 
 Other specialty 2 (0.4) 2 (0.9) — 
Elective status, n (%)   .97 
 Elective 150 (30.7) 73 (31.1) — 
 Urgent 335 (68.5) 162 (68.9) — 
 Not reported 4 (0.8) 0 (0) — 
Imaging, n (%)    
 Septations 94 (19.2) 69 (30.8) .003 
 Free fluid 149 (30.5) 79 (35.3) .34 
 Solid components 185 (37.8) 139 (62.1) <.001 
 Papillary projections 0 (0.0) 1 (0.5) .33 
 Ill-defined borders 8 (1.6) 3 (1.3) .33 
 Extension 4 (0.8) 6 (2.7) .002 
 Lymphadenopathy 8 (1.6) 4 (1.8) .28 
 Metastatic 2 (0.4) 3 (1.3) .11 
 Complex 129 (26.4) 82 (36.6) .01 
Size of mass, median (IQR), cm 6.7 (5.0–10.0) 9.7 (6.1–15.0) <.001 
Torsion noted intraoperatively, n (%) 80 (16.4) 72 (30.5) <.001 
Preoperative suspicion of malignancy, n (%) 28 (5.7) 41 (17.4) <.001 
Hospital, n (%)   <.001 
 1 43 (8.8) 15 (6.4) — 
 2 115 (23.5) 47 (19.2) — 
 3 40 (8.2) 8 (3.4) — 
 4 29 (5.9) 36 (15.3) — 
 5 15 (3.1) 7 (3.0) — 
 6 27 (5.5) 18 (7.6) — 
 7 61 (12.5) 27 (11.4) — 
 8 57 (11.7) 40 (17.0) — 
 9 30 (6.1) 14 (5.9) — 
 10 72 (14.7) 24 (10.2) — 

Other race includes Asian American, American Indian or Alaska Native, Hawaian Islander or Pacific Islander, multiracial, or self-reported other. —, not applicable.

Intraoperative torsion was identified in 21% of patients with a benign ovarian neoplasm. Among patients with benign neoplasms, there was no significant difference in the mass size of patients with and without torsion (9.54 vs 9.58 cm), but oophorectomy was more common in cases with torsion. Among the 152 patients with torsion, 47% underwent an oophorectomy.

Multivariable analyses revealed that torsion, imaging characteristics, preoperative suspicion for malignancy, and provider specialty were associated with increased odds of oophorectomy for benign lesions (Table 4). Patients with torsed ovaries or those who were found to have solid components on imaging had increased odds of oophorectomy (OR: 2.16, CI: 1.12–4.18; OR: 2.93, CI: 1.77–4.84, respectively). Each 1-cm increase in lesion size was associated with an 8% increased odds of oophorectomy (P < .0001). Preoperative suspicion of malignancy was associated with 2.7 times the odds of oophorectomy (P < .0001).

TABLE 4

Predictors of Oophorectomy in Benign Neoplasms in a Logistic Mixed-Effects Model

PredictorsOR95% Confidence LimitsP
Pediatric surgeona 2.68 1.12 6.41 .03 
Adult gynecologistsa 4.23 2.08 8.62 <.001 
Preoperative suspicion of malignancy 2.67 1.64 4.34 <.001 
Torsion observed intraoperatively 2.16 1.12 4.18 .02 
Solid components on imaging 2.93 1.77 4.84 <.001 
Size of mass, cm 1.08 1.05 1.10 <.001 
PredictorsOR95% Confidence LimitsP
Pediatric surgeona 2.68 1.12 6.41 .03 
Adult gynecologistsa 4.23 2.08 8.62 <.001 
Preoperative suspicion of malignancy 2.67 1.64 4.34 <.001 
Torsion observed intraoperatively 2.16 1.12 4.18 .02 
Solid components on imaging 2.93 1.77 4.84 <.001 
Size of mass, cm 1.08 1.05 1.10 <.001 
a

Reference group is pediatric and adolescent gynecologist (PAG). Other provider specialty includes urologists and unknown specialty.

Compared with PAGs, pediatric surgeons and adult gynecologists had significantly increased odds of performing oophorectomy (OR: 2.68, CI: 1.12–6.41; OR: 4.23, CI: 2.08–8.62, respectively). Further investigation revealed that case mix differed by provider specialty. Adult gynecologists treated older patients and more postmenarche patients, whereas pediatric surgeons treated more patients with torsion and performed more urgent cases (all P < .05). There was no significant change in the oophorectomy rates for benign disease across specialties on the basis of age or menarche status (P > .1 for both interactions). However, there were significant interactions between torsion and provider specialty (P = .006), as well as case status and provider specialty (P = .07) (Fig 1). For torsion and provider specialty, the rates of oophorectomy in cases with torsion were similar for pediatric surgeons and PAGs (43% vs 48%; P = .53), but in cases without torsion, pediatric surgeons had higher rates of oophorectomy compared with PAGs (32% vs 11%; P = .003) (Fig 1A). Rates of oophorectomy were significantly higher by adult gynecologists compared with PAGs in both cases with torsion (64% vs 48%; P = .03) and without torsion (41% vs 11%; P = .0001) (Fig 1A). For case status and provider, rates of oophorectomy in urgent cases were similar for pediatric surgeons and PAGs (32% vs 22%; P = .24), but in elective cases, pediatric surgeons had higher rates of oophorectomy compared with PAGs (43% vs 10%; P = .002) (Fig 1B). Rates of oophorectomy were significantly higher by adult gynecologists compared with PAGs in both urgent cases (46% vs 22%; P < .0001) and in elective cases (43% vs 10%; P = .0001) (Fig 1B).

FIGURE 1

Oophorectomy for benign disease differs on the basis of provider specialty and the presence of torsion (A, P = .006) and case status (B, P = .07). Pediatric surgeons and adult gynecologists had significantly higher oophorectomy rates than PAGs in cases without torsion (A) and in elective cases (B).

FIGURE 1

Oophorectomy for benign disease differs on the basis of provider specialty and the presence of torsion (A, P = .006) and case status (B, P = .07). Pediatric surgeons and adult gynecologists had significantly higher oophorectomy rates than PAGs in cases without torsion (A) and in elective cases (B).

With this multi-institutional study of 819 girls and women undergoing surgery for an ovarian neoplasm, we identified an overall malignancy rate of 11% but a 33% oophorectomy rate in patients with benign lesions across 10 children’s hospitals. This raises concerns that many patients may be undergoing an unnecessary oophorectomy for benign disease. The likelihood of undergoing an oophorectomy for benign lesions was independently associated with provider specialty, institution, premenarchal status, larger lesion size, and presence of solid components on preoperative imaging. Factors associated with a higher likelihood of a neoplasm being malignant included age, larger size, and a solid component or lymphadenopathy on imaging. The identification of these factors may allow for improved preoperative risk stratification and patient selection for ovary-sparing surgery versus oophorectomy.

The proportion of benign and malignant tumors identified in this study is similar to previous studies in which authors report 88% to 91% of pediatric ovarian neoplasms as benign.3,5,11,12 The identified higher risk for malignancy with younger age and premenarchal status is also consistent with previous studies. One previous report revealed that girls ages 1 to 8 years had the highest percentage of malignancy, and another revealed the majority of malignancies occurred in patients 9 to 14 years of age.3,11 The presence of solid components on imaging and increasing size were also independently associated with an increased risk of malignancy. These are consistent with previous studies that have identified size, imaging characteristics (including solid components), and abnormal tumor markers as factors associated with a higher likelihood of malignancy.2,3,11,12 

Several factors were independently associated with a higher likelihood of undergoing an oophorectomy for benign disease in our study. Both pediatric surgeons and adult gynecologists were more likely to perform an oophorectomy for benign disease as compared with a PAG. Investigation of case mix differences identified significant interactions between provider specialty and torsion and provider specialty and case status on the rate of oophorectomy for benign disease. As compared with PAGs, pediatric surgeons and adult gynecologists were more likely to perform an oophorectomy for benign disease in patients without torsion and in patients undergoing elective surgery. It is suggested in these data that the difference in oophorectomy rates for benign disease across provider specialty can be explained by differences in oophorectomy rates in the nonurgent cases and cases without torsion. These results are consistent with a previous administrative database study that revealed a higher likelihood of pediatric surgeons performing an oophorectomy for benign lesions than PAGs.6 The authors of that study also suggested that the presence of a PAG at a children’s hospital may result in increased use of ovary-sparing surgery for benign disease among all surgeons at an institution. One potential explanation for these differences may be the prioritization of the preservation of ovarian function by PAGs as compared to an emphasis of preventing missed malignancy by pediatric surgeons. The latter point may in part be supported by our finding that patients with a preoperative suspicion for malignancy were more likely to undergo an oophorectomy. The higher rate of oophorectomy performed by adult gynecologists may be due to the higher rate of malignancy in adult ovarian masses, which may lead them to perform a more aggressive procedure when operating on a pediatric patient with an ovarian mass.13 

Other identified factors associated with oophorectomy for benign disease may also be related to concerns for a higher risk of malignancy. Patients with a larger lesion, septations, complex features, extension into surrounding structures, and those with a solid component on imaging were also more likely to have an oophorectomy. Because this study and previous literature have revealed that premenarchal patients, larger masses, and evidence of a solid component, complex features, septations, or extensions into surrounding structures are all characteristics suspicious for malignancy, it is possible that an oophorectomy was performed in these patients out of concern for malignancy.11,14 Concerns for tumor spillage in patients with a suspected malignancy may also influence a surgeon’s decision to proceed with an oophorectomy because spillage can lead to additional treatments for patients. Our study revealed that tumor spillage occurred twice as often during ovary-sparing surgery but did not differ on the basis of provider specialty. Improved preoperative risk stratification may minimize the potential risks of spillage by allowing for better patient selection for ovary-sparing surgery.

Another clinical factor independently associated with higher odds of oophorectomy for benign disease was the concurrent diagnosis of torsion. This finding is concerning because the literature and specialty guidelines recommend against oophorectomy in the setting of torsion even if necrosis is present because the vast majority of patients are able to regain function of the ovary after detorsion.15,18 Although there may be concern for an occult malignancy within a mass, previous literature has revealed a 1.8% rate of malignancy in patients with torsion.19 However, in that same study, 88% of patients with ovarian torsion and a mass underwent an oophorectomy.19 In our study, 47% of patients with torsion and a mass underwent an oophorectomy. The higher oophorectomy rate in the setting of torsion may be due to concern for occult malignancy in the setting of an emergent operation that did not allow for further workup.

Although the importance of safely ruling out malignancy cannot be discounted, the impact of an oophorectomy on the fertility and overall health of the patient should also be considered. Patients who undergo a unilateral oophorectomy have been found to have a decreased ovarian reserve that puts them at a higher risk for having children with trisomy 21, similar to patients with advanced maternal age.20,21 In addition, there is an increased risk of premature ovarian failure and early menopause in patients who have undergone a unilateral oophorectomy.22 Another risk of unilateral oophorectomy is the possibility of contralateral disease and further surgery. Studies have revealed that as many as 23% of patients diagnosed with a mature cystic teratoma develop a neoplasm on the contralateral ovary requiring further surgery.23,26 If an oophorectomy was performed for the patient’s initial neoplasm, then preservation of as much ovarian tissue as possible must be considered during operations on the contralateral ovary, and there is the risk of surgical sterility if an oophorectomy is needed. Outside of their reproductive health, other studies have revealed an increased risk of cognitive impairment or dementia associated with unilateral (or bilateral) oophorectomy before menopause. Patients who underwent oophorectomy at a younger age were at an even higher risk for cognitive impairment or dementia.27 

All of these potential lifelong consequences of undergoing an oophorectomy underscore the need for improved preoperative risk stratification to minimize unnecessary oophorectomies in children.11 The identified variability in rates of oophorectomy for benign pediatric ovarian neoplasms and its association with surgeon specialty reveals the lack of clear guidelines for the management of ovarian masses in children. The greatest differences in oophorectomy rate for benign disease by provider type were during elective cases and cases without torsion. These data are encouraging because these are the cases in which preoperative risk stratification through additional testing (eg, imaging, tumor markers, etc) and multidisciplinary consultation can be applied to decrease unnecessary oophorectomies. We believe that children with ovarian masses will benefit from a multidisciplinary, standardized approach to care. Using the findings from this and other studies, our group of pediatric surgeons and PAGs has created a consensus algorithm for the workup and management of pediatric ovarian masses, which includes a multidisciplinary team discussion for any mass with concerning features (Fig 2). This algorithm is being prospectively evaluated across all institutions within our consortium. Results from this ongoing study will be used to calculate the sensitivity, specificity, and positive and negative predictive values of the algorithm and develop a scoring system for predicting the risk of malignancy preoperatively.

FIGURE 2

Algorithm for the management of pediatric ovarian masses. a Obtain tumor markers, including at least AFP, β–human chorionic gonadotropin, lactate dehydrogenase, cancer antigen 125, inhibin A, and inhibin B. b Pediatric surgery, pediatric gynecology, surgical oncology, +/− radiology, +/− oncology. c If indicated, consider staging with computed tomography scan and/or MRI of abdomen and/or pelvis, or computed tomography scan of the chest. d Consider peritoneal washings; inspection and/or biopsy of peritoneum, omentum, contralateral ovary, and retroperitoneal, and pelvic lymph nodes; or salpingo-oophorectomy if unable to obtain negative margins with oophorectomy alone.

FIGURE 2

Algorithm for the management of pediatric ovarian masses. a Obtain tumor markers, including at least AFP, β–human chorionic gonadotropin, lactate dehydrogenase, cancer antigen 125, inhibin A, and inhibin B. b Pediatric surgery, pediatric gynecology, surgical oncology, +/− radiology, +/− oncology. c If indicated, consider staging with computed tomography scan and/or MRI of abdomen and/or pelvis, or computed tomography scan of the chest. d Consider peritoneal washings; inspection and/or biopsy of peritoneum, omentum, contralateral ovary, and retroperitoneal, and pelvic lymph nodes; or salpingo-oophorectomy if unable to obtain negative margins with oophorectomy alone.

This study is not without limitations. The inconsistent measurement of tumor markers across sites limited our ability to investigate their association with malignancy. In addition, the retrospective design led to some missing and/or inconsistent data, including variability in the specific pathologic diagnoses and/or categories reported and the rate of unsuccessful attempts at ovary-sparing procedures across institutions. The nonrandom missingness of data limits our ability to use this data set to develop an unbiased scoring system for predicting the risk of malignancy preoperatively. Finally, this study was performed across children’s hospitals, so the results may not be generalizable to other institutions.

The vast majority of pediatric patients with ovarian masses have benign disease. However, a high number of these patients are still treated with oophorectomy. Factors associated with the choice of procedure and the risk of malignancy may allow for improved preoperative risk stratification and fewer unnecessary oophorectomies. Multidisciplinary approaches to the evaluation and treatment of pediatric ovarian masses are warranted.

Drs Lawrence, Onwuka, Minneci, and Deans conceptualized and designed the study, assisted in drafting the manuscript, assisted in statistical analysis, assisted in acquisition and interpretation of data, and reviewed and revised the manuscript; Drs Gonzalez, Aldrink, Fallat, Hertweck, Burns, Dillon, Ehrlich, Fraser, Grabowski, Hirschl, Kabre, Kohler, Lal, Landman, Leys, Mak, Sato, and Hewitt conceptualized and designed the study, performed substantial contributions to the acquisition of data, and critically reviewed the manuscript for important intellectual content; Drs Bence and Sujka and Ms Scannell performed substantial contributions to the acquisition of data and critically reviewed the manuscript for important intellectual content; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.

FUNDING: Supported by award UL1TR001070 from the National Center for Advancing Translational Sciences.

We acknowledge additional members of the Midwest Pediatric Surgery Consortium for their assistance with this project: Jacqueline M. Saito, Shawn D. St. Peter, Cynthia Downard, Michael Helmrath, Beth Rymeski, Kristine Corkum, Amin Afrazi, Brooks L. Rademacher, Manish T. Raiji, Brad Warner, Dan von Allmen, Gail Besner, R. Lawrence Moss, Keith Oldham, Marleta Reynolds, Jessica Kandel, Fred Rescorla.

     
  • AFP

    α-fetoprotein

  •  
  • CI

    confidence interval

  •  
  • IQR

    interquartile range

  •  
  • OR

    odds ratio

  •  
  • PAG

    pediatric and adolescent gynecologist

1
Skinner
MA
,
Schlatter
MG
,
Heifetz
SA
,
Grosfeld
JL
.
Ovarian neoplasms in children.
Arch Surg
.
1993
;
128
(
8
):
849
853; discussion 853–854
[PubMed]
2
Taskinen
S
,
Fagerholm
R
,
Lohi
J
,
Taskinen
M
.
Pediatric ovarian neoplastic tumors: incidence, age at presentation, tumor markers and outcome.
Acta Obstet Gynecol Scand
.
2015
;
94
(
4
):
425
429
[PubMed]
3
Papic
JC
,
Finnell
SM
,
Slaven
JE
,
Billmire
DF
,
Rescorla
FJ
,
Leys
CM
.
Predictors of ovarian malignancy in children: overcoming clinical barriers of ovarian preservation.
J Pediatr Surg
.
2014
;
49
(
1
):
144
147; discussion 147–148
[PubMed]
4
Cass
DL
,
Hawkins
E
,
Brandt
ML
, et al
.
Surgery for ovarian masses in infants, children, and adolescents: 102 consecutive patients treated in a 15-year period.
J Pediatr Surg
.
2001
;
36
(
5
):
693
699
[PubMed]
5
Zhang
M
,
Jiang
W
,
Li
G
,
Xu
C
.
Ovarian masses in children and adolescents - an analysis of 521 clinical cases.
J Pediatr Adolesc Gynecol
.
2014
;
27
(
3
):
e73
e77
[PubMed]
6
Gonzalez
DO
,
Cooper
JN
,
Aldrink
JH
, et al
.
Variability in surgical management of benign ovarian neoplasms in children.
J Pediatr Surg
.
2017
;
52
(
6
):
944
950
[PubMed]
7
Bristow
RE
,
Nugent
AC
,
Zahurak
ML
,
Khouzhami
V
,
Fox
HE
.
Impact of surgeon specialty on ovarian-conserving surgery in young females with an adnexal mass.
J Adolesc Health
.
2006
;
39
(
3
):
411
416
[PubMed]
8
Eskander
RN
,
Bristow
RE
,
Saenz
NC
,
Saenz
CC
.
A retrospective review of the effect of surgeon specialty on the management of 190 benign and malignant pediatric and adolescent adnexal masses.
J Pediatr Adolesc Gynecol
.
2011
;
24
(
5
):
282
285
[PubMed]
9
Hermans
AJ
,
Kluivers
KB
,
Wijnen
MH
,
Bulten
J
,
Massuger
LF
,
Coppus
SF
.
Diagnosis and treatment of adnexal masses in children and adolescents.
Obstet Gynecol
.
2015
;
125
(
3
):
611
615
[PubMed]
10
Aziz
D
,
Davis
V
,
Allen
L
,
Langer
JC
.
Ovarian torsion in children: is oophorectomy necessary?
J Pediatr Surg
.
2004
;
39
(
5
):
750
753
11
Oltmann
SC
,
Garcia
N
,
Barber
R
,
Huang
R
,
Hicks
B
,
Fischer
A
.
Can we preoperatively risk stratify ovarian masses for malignancy?
J Pediatr Surg
.
2010
;
45
(
1
):
130
134
[PubMed]
12
Rogers
EM
,
Casadiego Cubides
G
,
Lacy
J
,
Gerstle
JT
,
Kives
S
,
Allen
L
.
Preoperative risk stratification of adnexal masses: can we predict the optimal surgical management?
J Pediatr Adolesc Gynecol
.
2014
;
27
(
3
):
125
128
[PubMed]
13
Kinkel
K
,
Lu
Y
,
Mehdizade
A
,
Pelte
MF
,
Hricak
H
.
Indeterminate ovarian mass at US: incremental value of second imaging test for characterization-meta-analysis and Bayesian analysis.
Radiology
.
2005
;
236
(
1
):
85
94
[PubMed]
14
Amies Oelschlager
AM
,
Gow
KW
,
Morse
CB
,
Lara-Torre
E
.
Management of large ovarian neoplasms in pediatric and adolescent females.
J Pediatr Adolesc Gynecol
.
2016
;
29
(
2
):
88
94
[PubMed]
15
Cass
DL
.
Ovarian torsion.
Semin Pediatr Surg
.
2005
;
14
(
2
):
86
92
16
Dolgin
SE
,
Lublin
M
,
Shlasko
E
.
Maximizing ovarian salvage when treating idiopathic adnexal torsion.
J Pediatr Surg
.
2000
;
35
(
4
):
624
626
[PubMed]
17
American College of Obstetricians and Gynecologists’ Committee on Practice Bulletins—Gynecology
.
Practice Bulletin No. 174: evaluation and management of adnexal masses.
Obstet Gynecol
.
2016
;
128
(
5
):
e210
e226
[PubMed]
18
Oelsner
G
,
Cohen
SB
,
Soriano
D
,
Admon
D
,
Mashiach
S
,
Carp
H
.
Minimal surgery for the twisted ischaemic adnexa can preserve ovarian function.
Hum Reprod
.
2003
;
18
(
12
):
2599
2602
[PubMed]
19
Oltmann
SC
,
Fischer
A
,
Barber
R
,
Huang
R
,
Hicks
B
,
Garcia
N
.
Pediatric ovarian malignancy presenting as ovarian torsion: incidence and relevance.
J Pediatr Surg
.
2010
;
45
(
1
):
135
139
[PubMed]
20
Lass
A
.
The fertility potential of women with a single ovary.
Hum Reprod Update
.
1999
;
5
(
5
):
546
550
[PubMed]
21
Freeman
SB
,
Yang
Q
,
Allran
K
,
Taft
LF
,
Sherman
SL
.
Women with a reduced ovarian complement may have an increased risk for a child with Down syndrome.
Am J Hum Genet
.
2000
;
66
(
5
):
1680
1683
[PubMed]
22
Yasui
T
,
Hayashi
K
,
Mizunuma
H
, et al
.
Factors associated with premature ovarian failure, early menopause and earlier onset of menopause in Japanese women.
Maturitas
.
2012
;
72
(
3
):
249
255
[PubMed]
23
Chabaud-Williamson
M
,
Netchine
I
,
Fasola
S
, et al
.
Ovarian-sparing surgery for ovarian teratoma in children.
Pediatr Blood Cancer
.
2011
;
57
(
3
):
429
434
[PubMed]
24
Taskinen
S
,
Urtane
A
,
Fagerholm
R
,
Lohi
J
,
Taskinen
M
.
Metachronous benign ovarian tumors are not uncommon in children.
J Pediatr Surg
.
2014
;
49
(
4
):
543
545
[PubMed]
25
Templeman
CL
,
Hertweck
SP
,
Scheetz
JP
,
Perlman
SE
,
Fallat
ME
.
The management of mature cystic teratomas in children and adolescents: a retrospective analysis.
Hum Reprod
.
2000
;
15
(
12
):
2669
2672
[PubMed]
26
O’Neill
KE
,
Cooper
AR
.
The approach to ovarian dermoids in adolescents and young women.
J Pediatr Adolesc Gynecol
.
2011
;
24
(
3
):
176
180
27
Rocca
WA
,
Bower
JH
,
Maraganore
DM
, et al
.
Increased risk of cognitive impairment or dementia in women who underwent oophorectomy before menopause.
Neurology
.
2007
;
69
(
11
):
1074
1083
[PubMed]

Competing Interests

POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.

FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.

Supplementary data